History of Ancient Mathematics--Some Reflections on the State of the Art

Author(s): Sabetai Unguru

Source: Isis, Vol. 70, No. 4 (Dec., 1979), pp. 555-565
Published by: The University of Chicago Press on behalf of The History of Science Society
Stable URL: http://www.jstor.org/stable/230723 .
Accessed: 01/10/2013 10:52

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp

.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.

The University of Chicago Press and The History of Science Society are collaborating with JSTOR to digitize,
preserve and extend access to Isis.

http://www.jstor.org

This content downloaded from 129.72.2.27 on Tue, 1 Oct 2013 10:52:58 AM

All use subject to JSTOR Terms and Conditions
CRITIQUES & CONTENTIONS
History of Ancient Mathematics
SOME REFLECTIONS ON THE STATE
OF THE ART

By Sabetai Unguru*

Denn eben wo Begriffefehlen

Da stellt ein Wort zur rechtenZeit sich ein.
-JOHANN WOLFGANG VON GOETHE1

THE HISTORY OF MATHEMATICS typically has been written as if to illus-

trate the adage "anachronism is no vice." Most contemporary historians of
mathematics, being mathematicians by training, assume tacitly or explicitly that
mathematical entities reside in the world of Platonic ideas where they wait patiently
to be discovered by the genius of the working mathematician. Mathematical con-
cepts, constructive as well as computational, are seen as eternal, unchanging, unaf-
fected by the idiosyncratic features of the culture in which they appear, each one
clearly identifiable in its various historical occurrences, since these occurrences
represent different clothings of the same Platonic hypostasis.
Various forms of the same mathematical concept or operation are not considered
merely mathematically equivalent but also historically equivalent. Indeed mathemati-
cal equivalence is taken to represent historical equivalence. Since the mathematical
Forms are eternal and since in their works mathematicians of all ages share in the
expression of the same Forms, the specific mathematical idiom used by a mathemati-
cian has no bearing on the content of his thought. Mathematical language is at best a
secondary appurtenance of the mathematical culture of any epoch. The mathematical
kernel is untouched by the peculiar language used, since all mathematical languages
lead back to the same ideal Forms. This makes the various casts in which the same
mathematical truth has been expressed throughout the centuries completely equiva-
lent. As one of my colleagues put it: "Under such an ontology, the object of the
history of mathematics becomes the task of identifying the ideal forms present in the
work of each historical author and apportioning out proper credit to that mathemati-
cian who first gave expression to one of these eternal forms, i.e., who first brought it
out of the eternal Platonic realm into the world of human consciousness." This is
precisely the task performed traditionally by the historian of mathematics.
But if scholars continue to neglect the peculiar specificities of a given mathematical
*Departmentof the Historyof Science,Universityof Oklahoma,Norman,Oklahoma73019.
This paperis dedicatedto my parents,Zeidaand GhizaUnguru.In its presentformit owes a lot to the
criticismsof WillyHartnerand MatthiasSchramm,whom I thankfor theirassistance.I am also grateful
to the anonymous refereesand to many friends and colleaguesfor their critical suggestions.I am
exclusivelyresponsiblefor the views expressedhere.
'Faust I (Mephistophelesspeaks).
ISIS,1979,70 (No.254 4 555

This content downloaded from 129.72.2.27 on Tue, 1 Oct 2013 10:52:58 AM

All use subject to JSTOR Terms and Conditions
556 SABETAI UNGURU

culture, whether as a result of explicitly stated or implicitly taken-for-granted as-

sumptions, then by definition their work is ahistorical and should be recognized as
such by the community of historians. History, as Aristotle knew, focuses on the
idiosyncratic rather than the nomothetic.2 It is impossible for modern man to think
like an ancient Greek. Historical understanding, however, involves the attempt at
faithful reconstruction of the past. In intellectual history this necessarily means the
avoidance of conceptual pitfalls and interpretive anachronisms. Though it is impossi-
ble to think like Euclid, it is rather facile to think obtrusively unlike him. We cannot
know what went through Euclid's mind when he wrote the Elements. But we can
determine what Euclid could not have thought when he compiled his great work. He,
most likely, did not employ concepts or operations for which there is no genuine
evidence either in his time or in the works of his predecessors. This much is safe to
conclude. Furthermore, he clearly could not have foreseen what mathematicians and
historians of mathematics were going to do in the long run to his Elements; he could
not have used mathematical devices and procedures which were invented many
hundreds of years after his death. This much is obvious too. Given that we cannot
think like Euclid, we should, nevertheless, strive to avoid thinking unlike him when
elucidating and commenting on his writings. This is (and must remain) the historian's
goal. One way of thinking unlike Euclid is to use the algebraic approach in interpret-
ing his works.
It continues to be habitual among some historians of mathematics to say that what
really lies behind Euclid's geometrically couched statements are algebraic reasonings,
appearing in geometrical garb because of the lack of an appropriate algebraic
symbolism. This strikes me as both inaccurate and unilluminating. To see this, let us
ask the following question: how illuminating would it be to propose that Euclid really
thought in Sanskrit but because of his ignorance of the Sanskrit alphabet, had to use
the Greek one and consequently expressed himself in Greek? Greek mathematics
must be understood in its own right. This can be done by refusing to apply to its
analysis foreign, anachronistic criteria. The only acceptable meta-language for a
historically sympathetic investigation and comprehension of Greek mathematics
seems to be ordinary language, not algebra.
However, many scholars and in particular B. L. van der Waerden and Hans
Freudenthal do not endorse these ideas.3 Instead, they argue that mathematicians can
easily discern in Greek mathematics its underlying algebraic basis and interpret it
accordingly. Who would deny this manifest truth? The real question is: how accurate
and factual is the mathematicians' interpretation? Is the treatment to which mathe-
maticians have submitted Greek mathematics historically adequate? In fact, van der
Waerden and Freudenthal argue the cogency and the completeness of the mathemati-
cians' interpretation and treatment of Greek mathematics in algebraic form. But
there is a serious problem here. The fact that modern mathematicians can interpret
Greek mathematics algebraically is one thing. The conclusion that therefore the train
of thought of the Greek mathematicians was algebraic is an entirely different matter.
The step from the former statement to the latter is both a logical and a historical non
sequitur. "Logical," for obvious reasons; "historical," because it is possible to show

2Aristotle, Poetica 1451b 1-19.

3B. L. van der Waerden, "Defence of a 'Shocking' Point of View," Archive for History of Exact
Sciences, 1975, 15:199-210 and Hans Freudenthal, "What Is Algebra and What Has It Been in History,"
Arch. Hist. Exact Sci., 1977, 16:189-200, both written in response to my article "On the Need To Rewrite
the History of Greek Mathematics," Arch. Hist. Exact Sci., 1975, 15:67-114.

This content downloaded from 129.72.2.27 on Tue, 1 Oct 2013 10:52:58 AM

All use subject to JSTOR Terms and Conditions
 HISTORY OF ANCIENT MATHEMATICS 557

by an analysis of Greek mathematical texts that the assumption of an underlying

algebraic foundation for Greek mathematics leads to insoluble dilemmas and dread-
ful quandaries.4
In brief, van der Waerden defines algebra as "the art of handling algebraic
expressions like (a + b)2 and of solving equations like X2+ ax = b.''5 But no algebra
exists in Babylonian and pre-Diophantian Greek mathematical sources. "Babylonian
and Greek algebra" came into being only after the specific, numerical Babylonian
examples and the Greek geometrical propositions had been transcribed into algebraic
language; only as a result of the mathematician's elucidation of the texts was
"algebra" brought into existence.6 But the text itself did not present this elucidation;
the imaginative creation of the interpreter did so. And there is a mathematical
imagination and a historical imagination, and they typically run on different tracks.
Finally, continuing allegiance to the algebraic interpretation of Babylonian mathe-
matics is rendered still more untenable by the following rather damaging confession:
"In den eigentlich mathematischen Texten . . . die meisten Beispiele sicherlich von
ihrem Resultat aus hergerichtet sind."7 If this is true (and it seems to be, for most
answers are "nice," even numbers), then what is the basis for the claim that the
Babylonians solved equations? Do we as a rule solve equations by starting with the
answer?
Those who perceive an algebraic substructure bolstering Greek mathematics claim
that the Greeks started with algebraic problems but, then, translated them into a
geometric format. They have reached this conclusion, according to van der Waerden,
by studying "the wording of the theorems" and by trying "to reconstruct the original
ideas of the author. We found it evident that these theorems did not arise out of
geometrical problems [!]. We were not able to find any iAteresting geometrical
problem that would give rise to theorems like 11 1-4. On the other hand, we found
that the explanation of these theorems as arising from algebra worked well. There-
fore we adopted the latter explanation."8
But what evidence does van der Waerden present to demonstrate that "these
theorems did not arise out of geometrical problems"?The answer, he tells us, is that
no "interesting geometrical problem" leads to them. How does he know? Answer: he
could not find any. But the conclusion is unwarranted, since even if it is true that no
interesting geometrical problems led to them, it does not follow that noninteresting
geometrical problems did not lead to them either. Furthermore, what is an interesting
geometrical problem? Van der Waerden does not say, but the answer is implicit in
what follows: "we found that the explanation of these theorems as arising from
algebra worked well. Therefore... ." An interesting geometrical problem, then, seems
to be a problem the assumption of which "works well" in explaining the origin of the
theorems under discussion. And van der Waerden has decided arbitrarily (since he
could not check all possible geometrical theorems and problems) that there are no

4See "On the Need to Rewrite."

5"Defence," p. 199.
61n this context, S. Gandz, who advances his own algebraic interpretation of ancient mathematics in
"The Origin and Development of the Quadratic Equations in Babylonian, Greek and early Arabic
Algebra," Osiris, 1938, 3:405-557, says the following: "The commentary of NEUGEBAUER . . . has no
foundation in the text. It only shows how far away from the truth we may err, if we try, by all means, to
detect our modern school formulas in the old Babylonian text" (p. 423).
70tto Neugebauer, Vorgriechische Mathematik (Berlin: Springer-Verlag, 1934; reprinted Springer-
Verlag, 1969), p. 33.
8"Defence," pp. 203-204.

This content downloaded from 129.72.2.27 on Tue, 1 Oct 2013 10:52:58 AM

All use subject to JSTOR Terms and Conditions
558 SABETAI UNGURU

"interesting geometrical problems" working well under the circumstances. On the

other hand, what works well is the assumption of an underlying algebraic foundation
to Greek geometry. What does "working well" mean, then? Again, no answer is
provided, but it would clearly seem to mean something removing difficulties and
enabling one to cut through to the root and thus come up with "simple," "convinc-
ing," straightforward explanations. Ultimately, then, in the paragraph under discus-
sion, van der Waerden does say that Greek geometry (at least some important parts
of it) is, taken by itself, unfathomable, puzzling, weird, and that one can get rid of
these unsavory features by assuming a hidden algebraic basis to it. Therefore, "the
Greeks started with algebraic problems and translated them into geometric lan-
guage," Q.E.D.9
Leaving aside the circularity of the entire argument, and the conflation of logic and
history that it involves, van der Waerden's assertions represent an unconscious but
nevertheless clear-cut vindication of the argument that the real roots of the methodo-
logical position embodied in the concept "geometric algebra" lie in the modern
mathematician's ability to read geometric texts algebraically without any historical
qualms.

II

Why did the Greeks, according to the proponents of the idea, disguise algebra in
geometrical garb? Freudenthal gives three different answers: one historical, the second
philosophical (its pertinence entirely escapes me), and the third "traditional." His
"historical"answer speaks of a "torturouspath through foundations of mathematics,"10
which came to an end with the Eudoxian theory of proportions. But since there was
neither genuine foundational work nor a real Grundlagenkrisis (as Hasse and Scholz
and Van der Waerden referred to itl") in pre-Eudoxian times,'l2 speaking of "the
Greek end of the torturous path through foundations of mathematics, EUDOXOS'

91t is this very same approach which is involved in identifying the purely geometrical problem of the
application of areas as the Greek method of solution of quadratic equations, later equated with the
Babylonian method: the modern mathematician can indeed translate Greek geometry and Babylonian
specific-number manipulations into the algebraic language. Thus the parabolic application of areas, in
which one is asked to apply to a given straight line a rectangle equal to a given square, can be transcribed
as ax = b2, if the given line is a and the given square b2; to apply to the given line a rectangle equal to the
given square such that the applied rectanglefalls short of the second extremity of the given line by a square
(the elliptical application of areas) can be transcribed as x + y = a, xy = b2;and, to apply to the given line a
rectangle equal to the given square such that the applied rectangle exceeds the second extremity of the
given line by a square (the hyperbolic application of areas) can be transcribed as x - y a, xy = b2. This
mathematical possibility, however, is not a satisfactory historical justification for the claimed identity of
the Greek and the algebraic procedure. Moreover, strictly speaking, it is not the case that Euclid, Elements
1.44 corresponds exactly to the simple parabolic application of areas. The simple parabolic application
does not lead (as we saw) to a quadratic equation. If anything, it corresponds to the division of a given
product (area) by a given magnitude (line). Only Elements VI.28 and 29 lead, when transcribed algebrai-
cally, to complete quadratic equations corresponding respectively to elliptical and hyperbolic application
of areas. In this context, see A. Szabo, "Zum Problem der sog. 'Geometrischen Algebra' in Euklids
Elementen," completed in 1975 for a Festschrift in honor of Willy Hartner, p. 7 of the preprint.
1?Freudenthal, "What Is Algebra," p. 191.
"1H. Hasse and H. Scholz, "Die Grundlagenkrisis der griechischen Mathematik," Kant Studien, 1928,
33:4-34; B. L. van der Waerden, "Zenon und die Grundlagenkrise der griechischen Mathematik,"
Mathematische Annalen, 1940-1941, 117: 141-161.
12WilburR. Knorr, The Evolution of the Euclidean Elements: A Study of the Theory of Incommensu-
rable Magnitudes and Its Significance for Early Greek Geometry (Dordrecht/Boston: D. Reidel, 1975),
pp. 40-42, 50, 305-313; cf. also Hans Freudenthal, "Y avait-il une crise des fondements des mathematiques
dans l'antiquite?" Bulletin de la Societe Mathematique de Belgique, 1966, 8:43-55.

This content downloaded from 129.72.2.27 on Tue, 1 Oct 2013 10:52:58 AM

All use subject to JSTOR Terms and Conditions
 HISTORY OF ANCIENT MATHEMATICS 559

theory" is misleading and consequently Freudenthal's so-called historical answer is

nonhistorical and a nonan'swer. The other answers do not fare any better. For
example, in the "traditional" answer Freudenthal points out that "Once canonised,
the Elements were sacrosanct . . . The mathematical community was small. To be
understood within it, you had to quote EUCLID and to speak his language."13Fine, but
why did Euclid, then, adopt the very same language? In sum, Freudenthal's "three
main reasons" for the alleged disguise by the Greeks of algebra under the cloak of
geometry are not good reasons.
"The efficiency of a symbolism is determined by the ease with which the user can
move within it, by the algorithmic autonomy it provides."'4 Granted. But to judge
from the texts, neither the Babylonian nor the ancient Greek moved at all "within it,"
and it possessed for them no "algorithmic autonomy" whatever, since it was not yet
born. "It is the virtue of symbolism that it allows us for most of the time, rather than
identify object and symbol ... to forget about what the symbol means." But this task
is indeed impossible in Babylonian and Greek mathematical texts, where the object is
always identified, either as specific numbers or as spacial diagrams. It is impossible to
forget that "14, 30" means exactly "14, 30" and "line AB" means exactly "line AB"! It
is simply not true that "Almost never in the Elements or anywhere else in Greek
mathematics does AB mean a line or a line segment."'15
As Freudenthal would have it, Elements V is "algebra and nothing else"; it is,
moreover, "a general theory of magnitude . . . independent of dimension or any
characteristic of specific magnitudes."'16The problem with such a characterization is
the existence of Elements VII, in which many of the things dealt with in Book V are
repeated and applied specifically to numbers (integers). In the presence of a general
theory of magnitude, such a procedure would not have been just repetitious and
superfluous but outright senseless. Numbers, after all, are specific instances of
magnitude, and what is true of magnitudes in general is also true of numbers. In
writing Book VII, then, Euclid did not simply follow tradition (as Heath-thinks),'7
thereby merely proving over again for numbers propositions which he already proved
for all magnitudes, including numbers. Book V, it seems, presents a general theory of
proportion applicable to all kinds of magnitudes but not to numbers. The reason for
this is that numbers for the Greeks are not instances of a concept of general magni-
tude. "Magnitude, in fact, corresponds to one of the two divisions of quantity,
7roa6v, namely the continuous (as a line, a surface, or a body) whereas a number is
discrete."'8 Numbers (integers) are not illustrations of something else, they are
entities in their own right, with their own distinctive features, definitions, and so
forth. This is what enables Wilbur Knorr to say that "Book VII does not merely
duplicate Book V. It develops a body of analogous material for the separate class of
integers [my emphasis]; that is, it is required as an independent treatment, not a
duplication of a special case of Book V.'"19

13"WhatIs Algebra," p. 191. Freudenthal's philosophical answer reads: "Though in daily use by laymen
as well as mathematicians, fractions were taboo in highbrow mathematics, because philosophy forbade the
division of the unit" (ibid.).
14Ibid., p. 192.
'5Ibid.
16Ibid., p. 193.
17T. L. Heath, The Thirteen Books of Euclid's Elements, 3 vols. (Cambridge: Cambridge University
Press, 1908), Vol. II, p. 113.
18T. L. Heath, Mathematics in Aristotle (Oxford: Clarendon Press, 1949), p. 45.
19Knorr, Evolution of the Euclidean Elements, p. 309.

This content downloaded from 129.72.2.27 on Tue, 1 Oct 2013 10:52:58 AM

All use subject to JSTOR Terms and Conditions
 560 SABETAI UNGURU

Why should this be so? Basically, because of the Greek view that arithmetic is an
independent, not a derivative, discipline and that geometry and arithmetic are
different genera having their own domains, disposing of their own techniques of
demonstration, and dealing with their own subject matter. Pursuing them properly
means refraining from infringing upon the territory of one by means of the tools and
methods of the other.20
If I had been aware of the existence of Euclid's Data, argues Freudenthal, I "would
never have claimed there were no equations in Greek geometry." For Freudenthal,
the Data is a "textbook on solving equations." He summarizes the ninety-four
propositions contained therein in a succinctly and strikingly epigrammatic statement:
"Given certain magnitudes a, b, c and a relation F(a, b, c, x), then x, too, is given
.... "21 But the fact remains that Greek geometry contained no equations. One cannot
find even one equation in the entire text of the Data. Proof (as the Hindu mathemati-
cian would say): "Look!" Unless one has at his disposal the algebraic language and
the capacity to translate into it, it is impossible to sum up this little treatise of rather
varied content as offhandedly as Freudenthal has done. Indeed, had Euclid at his
disposal Freudenthal's functional notation, it is rather easy to infer that he would not
have needed ninety-four propositions to get his point across.
Each case in Euclid's Data is unique, having its own method of analysis, and none
is subsumable under or reducible to other cases. "Datarum magnitudinum ratio inter
se data est" (Prop. I) and "Si data magnitudo ad aliam magnitudinem rationem habet
datam, data est etiam illa magnitudine" (Prop. II)-to use perhaps the simplest
illustration possible-are not for Euclid both instances of "Given a, b, c, and y
F (a, b, c, x), x is also given," but two different problems, interesting in their own right,
having their own solutions. Of course, Freudenthal's description is mathematically
correct. Historically, however, it is wanting. Heath is much more to the point when
he says: "The Data ... are still concerned with elementary geometry, though forming
part of the introduction to higher analysis. Their form is that of propositions proving
that, if certain things in a figure are given (in magnitude, in species, etc.), something
else is given. The subject-matter is much the same as that of the planimetrical books
of the Elements, to which the Data are often supplementary."22
This is what the Data is, not a textbook on solving equations, but a treatise
presenting another approach to elementary geometry (other than that of the Ele-
ments, that is). Neither are Archimedes' works instances of "algebraic procedure in
Greek mathematics."23Heath's edition is "in modern notation."24It is faithful only to
the disembodied mathematical content of the Archimedean text, but not to its form.
And this is crucial. If one abandons Archimedes' form and transcribes his rhetorical
statements by means of algebraic symbols, manipulating and transforming the latter,
then clearly "the algebraic procedure" appears. But this procedure itself is not "in
Greek mathematics." It is a result (as Freudenthal himself states it) of "replacing
vernacular by artificial language, and numbering variables by cardinals, a quite
recent mathematical tool."25 Indeed! Archimedes' text is anchored securely in the
terrafirma of Greek geometry. If one is not willing to compress wording, to replace

20Aristotle, Posterior Analytics 75a37-75b20.

21"What Is Algebra," p. 194. By the way, references to the Data appear in my paper in a number of
places, e.g., p. 81, n. 26, and p. 108, n. 106.
22Heath, Elements, Vol. I, p. 8, my italics.
23"What Is Algebra," p. 195.
24T. L. Heath, The Works of Archimedes (Cambridge: Cambridge University Press, 1897), title page.
25"What Is Algebra," p. 196.

This content downloaded from 129.72.2.27 on Tue, 1 Oct 2013 10:52:58 AM

All use subject to JSTOR Terms and Conditions
 HISTORY OF ANCIENT MATHEMATICS 561

"vernacular" by artificial language, to introduce variables and number them by

cardinals, and to apply all the other technical tricks which are "quite recent mathe-
matical tools," then Archimedes' proof of Proposition X of HEPI EAIKfQN is
geometric, not algebraic. This was discerned in a curious way even by Heath, who
justified his algebraic procedure and the use of the symbols A1, A2, . . An, "in order
to exhibit the geometrical character of the proof."26
Dijksterhuis himself in his Archimedes said: "in a representation of Greek proofs in
the symbolism of modern algebra it is often precisely the most characteristic qualities
of the classical argument which are lost, so that the reader is not sufficiently obliged
to enter into the train of thought of the original. . . . s27

III

Both Freudenthal and van der Waerden have constructed identical operative defini-
tions of algebra, thereby creating significant problems in their analyses of Greek
geometry. Freudenthal says: "This ability to describe relations and solving proce-
dures, and the techniques involved in a general way, is in my view of algebra such an
important feature of algebraic thinking that I am willing to extend the name 'algebra'
to it. . . . But what is in a name?"28However, it is precisely the inability of the
Babylonian mathematician "to describe relations and solving procedures, and the
techniques involved in a general way" that warrants his disqualification as algebraist.
What the Babylonian mathematician lacks is precisely the ability to dispense with
specific, definite numbers, and it is this deficiency that dictates the particular form of
his approach. What he can produce is recipes, not general formulas.
With respect to the Greek mathematician (geometer), on the other hand, though it
is legitimate to see his approach as a general approach (the so-called theorem of
Pythagoras is true of any right-angled triangle, etc.), the language he uses is the
geometric language and the generality involved is an outgrowth of dealing with
geometrical and not with algebraic entities. Consequently, by Freudenthal's own
criteria of "algebraic thinking," Babylonian and Greek mathematics are nonalgebraic.
"What's in a name?"asks Freudenthal and uses the question even as a motto for his
article. The answer, clearly enough, is "it depends." Names are words, and words are
important when used thoughtfully. As a matter of fact, it is possible to argue that all
there is is, one way or another, in words. The Iliad and Hamlet are in words; and so is
the Magna Charta. The Bible is in words; and so is the American Declaration of
Independence. All of mathematics is in a very definite sense in words. Thought and
feeling (beyond inarticulate physiological reactions) are in words. Artistic experience
is in a proper sense in words, for no informed, thoughtful reaction to and communi-
cation about a work of art is possible in the absence of articulate expression, which is
again in words. Our meaningful access to reality (whatever it may consist of) is
always mediate: we know the world through words.29

26Heath, Archimedes, p. 109, my italics.

27E. J. Dijksterhuis, Archimedes (Copenhagen: Munksgaard, 1956), p. 7.
28"What Is Algebra," pp. 193-194, my italics.
291n a pregnant way, the Venerable Inceptor expressed this view as follows: "Si dicas: nolo loqui de
vocibus sed tantum de rebus, dico quod quamvis velis loqui tantum de rebus, tamen hoc non est possibile
nisi mediantibus vocibus vel conceptibus vel aliis signis" (William of Ockham, Commentary on the
Sentences, I, dist. 2, quest. 1, in Super quatuor libros sententiarum (In sententiarum I), being Vol. III
(1495) of Guillelmus de Occam, O.F.M., Opera plurima (Lyons, 1494-1496; reprinted London: Gregg
Press, 1962, in 4 vols.).

This content downloaded from 129.72.2.27 on Tue, 1 Oct 2013 10:52:58 AM

All use subject to JSTOR Terms and Conditions
562 SABETAI UNGURU

But words can be misused. Mein Kampf and the Protocols of the Elders of Zion
are also in words; and so is The National Enquirer. Words are powerful weapons,
and men are governed (or misgoverned) with, by, and in words. And so, "what's in a
word?" As always, "it all depends." Is the word used carefully? Does the user follow
the advice of Paul to the Ephesians: "Let no man deceive . . . with vain words"? Are
words used and understood pertinently with reference to the subject matter, accord-
ing to the old legal maxim, "Verba accipienda sunt secundum subjectam materiam?"
If the answer to the above questions is positive, then there is a lot in a word; if
negative, the word is misleading and therefore dangerous.
The use of the word "algebra" as a term descriptive of Babylonian and Greek
mathematics is a misuse of the word. When the questions enumerated above are
asked in connection with that use, all the answers come out negative. The word
'"algebra"is used carelessly; its use is deceiving since it leads to a translation of
ancient mathematical texts into a historically inappropriate language; and, if "al-
gebra" has its proper meaning, the use of the term is unsuited to the subject matter.
Words are judgments, or, as Nietzsche put it, preconceived judgments; and this is
how it should be. But some judgments carry conviction while others are blatantly
unjust. The word "algebra" in the context discussed belongs to the latter category.
Enthusiasts of algebraic interpretations of Greek geometry have violated one of the
fundamental tenets of historical scholarship. History is the study of the present traces
of past events from the standpoint of change and the particular, the idiosyncratic.30
Although long-lasting structures, stable frameworks, and durable, quasi-constant
features are legitimate topics of historical investigation, they are not what makes
history what it is.31 History is primarily, essentially interested in the event qua
particular event, in the specific happening, in change from an identifiable, individual
characteristic to another identifiable, individual characteristic. History is not (or is
primarily not) striving to bunch events together, to crowd them under the same
heading by draining them of their individualities. On the contrary, history is the
attempt at understanding each past event in its own right. The domain of history,
then, is the idiosyncratic.
The historian of ideas does not discharge his obligation by showing merely the
extent to which past ideas are like modern ideas. His main effort should be in the
direction of showing the extent to which past ideas were unlike modern ones,
irrespective of the fact that they might (or might not) have led to the modern ideas.
This is a wise methodological tack, since it enables the historian to avoid reductive
anachronism while channeling his historical empathy toward an understanding of the
past in its own right. It is also wise to take the written documents of the past to mean
precisely what they say, short of clear-cut proof to the contrary. There is no historical
advantage whatever growing out of the gratuitous assumption that the men of old
played tricks on us by systematically hiding their line of thought.
I shall not presume to define here what mathematics is, as that is best left to
mathematicians. Besides, there are plenty of definitions available, running the gamut
from Bertrand Russell's to Nicolas Bourbaki's.32Every reader can easily take his

30See G. R. Elton, The Practice of History (Sidney: Sidney University Press, 1967), pp. 8-12.
31". .. there is more to history than the study of persistent structures and the slow progress of evolution"
(Fernand Braudel, The Mediterranean and the Mediterranean World in the Age of Philip II, 2 vols., New
York: Harper & Row, 1975, Vol. II, p. 901).
32". . mathematics may be defined as the subject in which we never know what we are talking about, nor
whether what we are saying is true" (Bertrand Russell, Mysticism and Logic and Other Essays, New York:

This content downloaded from 129.72.2.27 on Tue, 1 Oct 2013 10:52:58 AM

All use subject to JSTOR Terms and Conditions
 HISTORY OF ANCIENT MATHEMATICS 563

pick. But I can say safely what mathematics is not. It is certainly not history. The
domain of mathematics is not the idiosyncratic, but, in a very real sense, the
nomothetic, since what mathematicians do is to show that from certain assumptions
about as yet unidentified objects some conclusions about the same objects will follow
necessarily, by rule.
The history of mathematics is history not mathematics. It is the study of the
idiosyncratic aspects of the activity of mathematicians who themselves are engaged in
the study of the nomothetic, that is, of what is the case by law. If one is to write the
history of mathematics, and not the mathematics of history, the writer must be
careful not to substitute the nomothetic for the idiosyncratic, that is, not to deal with
past mathematics as if mathematics had no past beyond trivial differences in the
outward appearance of what is basically an unchangeable hard-core content.
In mathematics (like in anything else) form and content are not independent
variables. On the contrary, they mutually condition one another and neither is
immune to change. A certain form permits only a certain content, and a new content
requires a new form. This is why the methodological approach which casts indiscrim-
inately the algebraic shadow over the garden of Greek mathematics obscures pre-
cisely those features which make it Greek mathematics. Instead of showing the degree
to which it was unlike modern, post-Renaissance mathematics, that approach, by
greatly overemphasizing the similarities, prevents an understanding of Greek mathe-
matics in its own right. It also leads in the long run to the untenable view that the
Greek mathematicians did not mean what they said, but that they hid "admirably"33
their line of thought. Coupled with this is the great danger of easily "discerning"
problematic or nonexistent influences between mathematical cultures a world apart,
simply because when submitted to the algebraic cure all mathematical cultures look
alike.
Entrenched as it is, the traditional interpretation of the history of ancient mathe-
matics must give way to a new, more sympathetic, and historically responsive
interpretation, simply because the old interpretation has outlived its usefulness and is
now an obstacle on the road to a sensitive historical understanding of ancient
mathematical texts. After all, like scientific theories, historical theories are tentative
attempts to make sense of the past; they are provisional by their very nature, and
consequently their authors should not be dreaming hopelessly of endowing them, in
God-like fashion, with eternal life and immaculate beatitude.
Otto Neugebauer is right. Speaking of the fact that it was the Hindus and not the
Babylonians who introduced a sign for zero to be used always whenever required in
the writing of numbers, Neugebauer makes the following pertinent remark:

It seems to me . . . that the awarenessof the arbitrarinessand the purelyconventional,

symbolic characterof all means of expression does not arise in the frameworkof a
continuoushistoricaldevelopment,which, of course, rests on the direct traditionfrom
generationto generation;the awarenessof all thesethingsbecomesabsolutizedinto fixed
and rigidformswhichcannotbe substantiallychangedof one'sown accord,as this largely
transcendsthe analyticalcapacityof mankind.Only men who are the heirs of an alto-

Barnes and Noble, 1971, pp. 59-60) and "A mathematical theory ... contains rules which allow us to assert
that certain assemblies of signs are terms or relations of the theory, and other rules which allow us to assert
that certain assemblies are theorems of the theory" (Nicolas Bourbaki, Elements of Mathematics: Theory
of Sets, Paris/London: Addison-Wesley, 1968, p. 16).
33B. L. van der Waerden, Science Awakening (Groningen: P. Noordhoff, 1954), p. 172.

This content downloaded from 129.72.2.27 on Tue, 1 Oct 2013 10:52:58 AM

All use subject to JSTOR Terms and Conditions
564 SABETAI UNGURU

getherdifferenthistoricaltraditionare able to use freelythe foreignmeansof expression

and to recognizeboth their limits and their potentialities.34

Now this seems indeed to be the case with respect to the introduction of the algebraic
approach by Viete, Fermat, and Descartes, men of genius belonging to another
culture than the Greek, but who managed somehow to discern in what the Greeks
had done (geometry) precisely what the Greeks themselves never dreamt about when
they were doing it, namely a hidden algebraic structure, which the moderns set about
to extract from the Greek texts. This is the true historical origin of the concept
"geometrical algebra." It is the intellectual product of foreigners, barbarians, reading
Greek mathematical texts in light of their own idiosyncrasies, their own barbarian
approach, and "seeing" in it what the Greeks, the autochthons, never put into it,
namely, an algebraic train of thought. Mutatis mutandis, like the Babylonians with
the general concept of zero, the Greeks never came up with a symbolic approach; it
remained for the sixteenth- and seventeenth-century Europeans (playing somewhat
the role of the Hindus in our comparison), the heirs and at the same time the usurpers
of the Greeks, to invent the general symbolic approach and thereby to "perceive"its
roots within the confines of Greek geometry.
Though what one calls a thing is, to begin with, merely a convention, once the
calling (naming) has been socially accepted, departures from the standard usage
without further ado are misleading and can be dangerous. Whatever algebra might
"really" be, the term as standardly used means something definite, as do most of the
words used in common parlance. This is what makes communication possible. A
"table" is a table. A "chair"is a chair. Even Freudenthal agrees with that, since he
says: "'algebra' has a meaning in everyday language just as 'chair'and 'table' have."35
Calling, then, a tree "table"is misleading, in spite of the fact that trees can (and quite
often do) become tables. (As a matter of fact-and this is crucial-quite often they do
not.) By the same token, calling a tree "chair" is misleading. In such an arbitrary
naming procedure, one substitutes one of the many potentialities of the object for its
reality. This is dangerous, since trees are potentially not just tables or chairs, but also
coffins or houses. Calling a tree "table," then, is misleading not only because it takes
the potential for the real but also because it neglects all but one of the various
potentialities of the object.
Precisely as it is only hindsight that enables one to call legitimately a certain tree
"table," or "chair," or "coffin," it is only unwarranted historical hindsight that has
enabled scholars to call Greek geometry "algebra," by setting up just one of the
potentialities of Greek geometry into a chosen entelechy. There may exist, by divine
decree, a chosen people. However, "chosen" entelechies, in the perfectly natural case
of multiple potentialities, are post factum creations of the mind of the historian-
philosopher running rampant, since the whole historical point consists exactly in the
necessity to show that in the actual historical process only "the chosen entelechy" has
been realized.

34 Vorgriechische Mathematik, p. 78: "Mir scheint . . . dass im Rahmen einer kontinuierlichen geschicht-
lichen Entwicklung, die ja auf der direkten Tradition von Generation zu Generation beruht, das Bewusst-
sein der Willkiirlichkeit und des rein konventionellen symbolischen Charakters aller Ausdrucksmittel gar
nicht ensteht, dass alle diese Dinge zu aboluten und gegebenen Formen werden, die aus freien Stiicken
wesentlich abzuandern das analytische Verm6gen der Menschen weit iibersteigt. Erst Menschen die selbst
einer ganz anderen geschichtlichen Tradition entstammen, sind imstande, die fremden Ausdrucksmittel
frei zu gebrauchen und ihre Schranken wie ihre Moglichkeit zu erkennen."
35"WhatIs Algebra," p. 193.

This content downloaded from 129.72.2.27 on Tue, 1 Oct 2013 10:52:58 AM

All use subject to JSTOR Terms and Conditions
 HISTORY OF ANCIENT MATHEMATICS 565

It is true that names are conventions. But conventions fulfill a very important
function, making articulate communication (i.e., intelligent life) possible. Abiding by
them enables one to carry on in everyday life. Blatant transgressions against socially
accepted conventions, on the other hand, prevent normal communication and can be
rather troublesome. It is mere convention to kiss and embrace one's bride at the
wedding. Refuse to do it, "because it is a mere convention," and you are in for some
real trouble.
The name "algebra," like all names, is a convention (although it has some very
definite historical roots). But it means something recognizable in common parlance.
Apply it indiscriminately to what is obviously geometry and you have not merely
breached a useful convention, you have thereby created a new one, less definite,
sharp, and useful than the one you violated, since it substitutes potentiality for
reality. And although this is possible, it is wrong historically, since history deals with
reality (what happened) and not with potentiality (what could have happened
logically). The approach of Freudenthal, van der Waerden, and their cohort
substitutes logic for history.36

36Here, I have in mind Andre Weil's unprecedented missive to the editor of the Archivefor History of
Exact Sciences, entirely repetitive in its few non ad hominem passages of the arguments of van der
Waerden and Freudenthal: "Who Betrayed Euclid?" Arch. Hist. Exact Sci., 1978, 19:91-93. Concerning
this letter, the less said the better. In adopting this position, I am guided by Simone Weil's words in her
sensitive and penetrating essay on the Iliad (The Iliad or the Poem of Force, Wallingford, Pa.: Pendle Hill,
n.d., pp. 3, 36): "To define force-it is that x that turns anybody who is subjected to it into a thing.
Exercised to the limit, it turns man into a thing in the most literal sense: it makes a corpse out of him.
Somebody was here, and the next minute there is nobody here at all;" And: "The man who does not wear
the armor of the lie cannot experience force without being touched by it to the very soul. Grace can prevent
this touch from corrupting him, but it cannot spare him the wound."

METHODOLOGY, PHILOLOGY, AND PHILOSOPHY

By Wilbur Knorr*

IN A NOTE to his article "The Philosophical Sense of Theaetetus' Mathematics"

(Isis, 1978, 69:489-513, n. 88), M. F. Burnyeat raises some ostensibly damaging
observations against the argument in my book The Evolution of the Euclidean
Elements (Dordrecht: Reidel, 1975). I would like to make a few remarks in the hope
of correcting at once a number of basic misconceptions.
A major issue seems to center on the interpretation of a single sentence in Plato's
Theaetetus: is it to be rendered "in this case [i.e., 17] he [Theodorus] for some reason
came to a standstill [enescheto]," as most commentators have chosen; or, alterna-
tively, "in this case for some reason he encountered difficulty," as I have taken it,
supported by the textual analysis given by R. Hackforth (Mnemosyne, 1957, 10:128;
*Institute for Advanced Study, Princeton, N.J. 08540
ISIS, 1979, 70 (No. 254)

This content downloaded from 129.72.2.27 on Tue, 1 Oct 2013 10:52:58 AM

All use subject to JSTOR Terms and Conditions